What is the windward problem?

The windward problem is when the goal you want sits "dead upwind"—directly against the prevailing force of people's incentives, habits, or trust—so steering straight at it stalls you, the way a sailboat pointed into the wind goes "in irons" and stops. The goal is still reachable; the head-on heading is the one approach that can't work. The fix is to tack: reach it through a series of deliberate, angled moves rather than a frontal push.

Why do AI adoption mandates fail?

Because adoption is a bottom-up decision being forced top-down. A 2026 WalkMe survey found roughly eight in 10 workers are avoiding or rejecting their company's AI tools, often doing the work manually instead. Mandating use against a trust gap or a capacity crunch compresses resistance rather than reducing it—people build a "rock wall." Tools that genuinely help get adopted with no memo; tools pushed against the grain stall regardless of how forceful the mandate is.

How do you get a team to adopt AI without forcing it?

Tack instead of pushing. Name the real headwind (trust, capacity, a past failed mandate), then make the smallest angled move that produces a visible win—a volunteer pilot, one high-value opt-in task, one workflow redesigned end to end. Convert each win into the next leg, let managers carry it rather than executives, and measure actual outcomes instead of logins. McKinsey's AI high performers got there by redesigning workflows, not by mandating usage.

Isn't top-down change sometimes necessary?

Yes. When the goal aligns with what people already want—downwind or across the wind—going straight is the fastest, most honest path, and tacking would just be slow. Direct authority is the right tool when little is resisting it. The windward problem only applies when the goal runs against incentives, trust, or capacity. The skill is diagnosing the wind direction first, then choosing a direct push or an indirect tack accordingly.

What's the difference between top-down and bottom-up adoption?

Top-down adoption is driven by mandate—leadership decides, then requires compliance. Bottom-up adoption is driven by self-interest—individuals pick up a tool because it visibly makes their own work better, the way developers adopted IDEs with no one ordering them to. The windward framing maps this to wind direction: bottom-up goals sit downwind (sail straight), while top-down behavior change against the grain sits upwind (you must tack, not force).

The Windward Problem: Why Forcing AI Adoption Stalls

Point a sailboat straight at the wind and it stops. The sails stop pulling and start flapping, the boat loses steerage, and you sit there—bow aimed exactly where you want to go, going nowhere. Sailors have a name for this dead, embarrassing state: the boat is "in irons." The harder the wind blows, the more stuck you are. The one direction a boat cannot sail is the one direction it's pointed.

Most leaders are trying to sail straight into the wind right now, and calling it an AI strategy. The thrust is enormous and the boat isn't moving. A 2026 WalkMe survey of 3,750 executives and employees found that "more 54% of workers bypassed their company's AI tools in the past 30 days and completed the work manually instead," and that "roughly eight in 10 enterprise workers are either avoiding or actively rejecting the technology" their companies are spending record sums to deploy. Maximum push, zero progress. That's the windward problem: when the goal you want sits dead upwind—against the prevailing force of people's incentives and habits—and steering straight at it just puts you in irons.

This is a how-to, not a lament. We'll cover what the windward problem actually is, why pushing harder makes it worse, the proof that your destination is fine but your heading is wrong, and the five-move maneuver sailors use to reach a point they can't sail toward directly: tacking. By the end you'll know when to go straight and when to stop fighting the wind.

What the windward problem actually is

Here's the part almost everyone gets wrong about a sailboat. We picture the wind pushing the boat like a hand on a toy, so "obviously" you can't go toward the wind and the boat just drifts wherever the air shoves it. That's not how it works. A sail is a wing. "In points of sail that range from close-hauled to a broad reach, sails act substantially like a wing, with lift predominantly propelling the craft"—the same aerodynamic lift that holds up an airplane. That's why a boat can sail upwind, diagonally across the wind, which looks impossible until you understand the physics.

But there's a hard limit. A sail can't make lift if you point too close to the wind. "A sailing craft cannot sail directly into the wind, nor on a course that is too close to the direction from which the wind is blowing, because the sails cannot generate lift in this no-go zone." That no-go zone is an arc of about 45° on either side of dead upwind—roughly a quarter of the compass that is simply forbidden. Sail into it and you go in irons.

So the windward problem isn't "this goal is impossible." Windward goals are completely reachable. It's that there's one specific way you can't approach them: head-on. And the comforting management instinct—see the target, point at the target, apply force—aims you straight into the no-go zone. Call it a model, not a law of physics: people aren't wind. But the shape holds eerily well. Some goals sit downwind of what your team already wants, and you can run straight at them. Others sit dead to windward of entrenched incentives, and the direct line is the one heading guaranteed to stall.

Why pushing harder makes it worse

Let's steelman directness first, because it's usually right. When the wind is at your back—when the goal lines up with what people already want—going straight is not just fine, it's fastest. Nobody needed a mandate to make developers adopt better autocomplete or faster laptops. Force is the correct tool when nothing is resisting it. The mistake is assuming that because directness works downwind, it works in every direction.

It doesn't, and the failure mode is specific: the harder you push into the wind, the harder the wind pushes back. One engineer described a hospital's attempt to force a process change in terms a sailor would recognize: "The harder you push, the more they'll form a rock wall against your demands... In order to have surgeons participate you need to earn their trust." The strategist B.H. Liddell Hart built an entire theory on this, the strategy of the indirect approach: "A direct approach to the object exhausts the attacker and hardens the resistance by compression, whereas an indirect approach loosens the defender's hold by upsetting his balance." Compression. The frontal push doesn't just fail to move the obstacle; it manufactures a stronger one.

This is the core of why AI adoption mandates fail. As one developer put it on Hacker News, "management is trying to top-down force adoption of something that operates at the individual level and whose adoption is thus inherently a bottom-up decision... We never saw CEOs issue memos 'reorganizing' the company around IDEs or software frameworks and mandate that the employees use them, because employees naturally saw their productivity gains and adopted them organically." The whole problem of top-down vs bottom-up adoption is a windward problem in disguise: tools that genuinely help get adopted downwind, with no memo; tools pushed against the grain sit in irons no matter how loud the mandate.

We have a clean precedent. The return-to-office push was the same head-on maneuver, and it stalled the same way—one in five HR professionals admitted their in-office policy "was meant to make staff quit," which is what happens when you confuse the noise of compression for progress (more on that in our breakdown of the RTO mandate backlash). And the headwind is stiffening: Microsoft's 2025 Work Trend Index found 53% of leaders say productivity must increase, but 80% of the global workforce say they're lacking enough time or energy to do their work. Pushing a new mandate into an out-of-capacity crew is sailing dead into a gale and wondering why the boat won't move.

The destination is fine—your heading is wrong

The most important thing about the windward problem is what it tells you not to do: don't abandon the goal. The wind isn't telling you AI is a bad destination. It's telling you your heading is. A windward goal is still a goal—it just has a forbidden approach.

The data makes this almost literal. McKinsey's 2025 State of AI found 88% of organizations now use AI in at least one function—nearly everyone has the boat in the water—yet only about 6% qualify as high performers actually capturing value. The difference between the stalled fleet and the few who are moving isn't more thrust. It's heading: McKinsey's high performers are far more likely to have fundamentally redesigned their workflows (55% versus 20% of everyone else). They didn't shout louder at the wind. They changed their angle to it.

You can hear the in-irons groan from the teams that didn't. One engineer described his company "replacing an already-working system with a half-broken AI one just so we can say we use AI, and then forcing everyone at the company to use it just so we can release a press release saying 'all of our developers use AI.'" That's a boat pinned in the no-go zone, sails flapping, while leadership celebrates the heading. The goal was never the problem. The straight line to it was.

How to tack toward a windward goal

Sailors reach a windward mark by beating to windward—"in a zig-zag fashion with a series of tacking maneuvers," sailing at an angle the boat can hold, then switching sides, then switching back, until the indirect path adds up to the place you couldn't point at. In plain terms: you win a forced-adoption goal through a series of smaller, indirect wins, not one top-down push. Here's how to run the same maneuver on change management resistance.

Name the true wind

You can't trim for a wind you haven't located. Before you push anything, name the force actually opposing the goal—not the people, the force. Is it a trust gap? WalkMe found only 9% of workers trust AI for complex decisions versus 61% of executives—a "52-point trust chasm." Is it capacity, like Microsoft's 80% running on empty? Is it a credibility hangover from the last mandate? The direction of the wind decides every other move. Pointing at the goal without reading the wind is how you sail confidently into irons.

Bear off to your best angle

The rookie error is "pinching"—pointing as close to the goal as you possibly can, which kills your speed and stalls you. Don't pinch. Bear off to the angle where the sail actually fills: the smallest first step that produces a real, visible win against the headwind. A volunteer pilot team. One high-value, opt-in task where AI obviously beats the manual way. A single workflow redesigned end to end. It will feel indirect—you're not aimed straight at "everyone uses AI"—and that's the point. An angled course that moves beats a direct course that doesn't.

Tack, then come about

One tack isn't the trip. You make progress on one heading—a win on this team, this use case—then come about and convert that proof into the next leg. Each tack should make the next one easier, because the people who won on the last leg become the crew that pulls the next. This is where managers, not mandates, do the work: the manager-as-multiplier effect on AI adoption is real, and it's the difference between a leader shouting the heading and a crew that knows how to sail it. Tacking is getting buy-in for change one defensible leg at a time, not in one impossible jump.

Measure VMG, not your heading

Racers don't measure how close their bow points to the mark—they measure "velocity made good," actual progress toward it. A boat pointed dead at the goal with no speed has zero VMG. A boat angled away but moving has plenty. Most adoption dashboards measure heading: seats provisioned, logins, "all developers use AI." Those are the bow angle, not the progress. Measure VMG—work that's genuinely faster or better—which means managing outcomes instead of mandating the behavior. Usage theater is the most expensive way to look like you're moving while in irons.

Build a boat that points higher

A better boat needs a smaller angle—it can sail closer to the wind and still draw power. The same is true of your tools. The reason the IDE never needed a memo is that it pointed so close to what people already wanted that adoption was downwind. Reduce the angle by reducing friction: when the work lives in one place instead of scattered across a dozen tabs (the very cost of context-switching that makes people reject one more tool), the headwind drops. This is the bet behind Coommit—video, a shared canvas, and contextual AI in one surface, so the AI shows up inside the work people are already doing rather than as a separate destination they have to be forced to visit. A tool people want points higher into the wind on its own.

When to sail straight—and when to tack

Tacking is not always right; sometimes the direct line is correct and zig-zagging is just slow. So run one test before you choose: which way is the wind? If the goal aligns with what people already want—if it's downwind or across the wind—sail straight at it. Directness is the fastest path when nothing is resisting, and turning that into a committee is its own waste, the cousin of trying to fix a coordination problem with more headcount (Brooks's law in reverse). But if the goal runs against incentives, trust, or capacity—if it sits to windward, in the no-go zone—stop pushing. That is the windward problem, and force is the one tool that can't solve it. The frontal assault that feels like leadership is the one heading that guarantees you stay in irons. Read the wind first; pick the maneuver second.

The longest way round

Every instinct in a results-driven culture says the shortest distance to a goal is a straight line, and to treat any indirect path as a detour for people who lack conviction. On open water with the wind behind you, that instinct is right. But the moment your goal sits dead upwind, "go straight at it" quietly becomes "go nowhere, loudly"—and the harder you push, the more the resistance hardens by compression. Liddell Hart had the whole thing in nine words: "In strategy, the longest way round is often the shortest way home."

The AI era is going to make more of your goals windward, not fewer—every tool gets faster while human trust, capacity, and habit stay exactly as stubborn as they've always been. The teams that win the next two years won't be the ones with the most thrust or the loudest mandate. They'll be the ones who can read the wind and tack. Pick the one upwind goal your team is in irons on right now, find the smallest angled move that actually fills the sail, and bear off. You'll reach it—just not in a straight line.